John Benjamins Publishing Company
Total Page:16
File Type:pdf, Size:1020Kb
John Benjamins Publishing Company This is a contribution from Gesture 18:1 © 2019. John Benjamins Publishing Company This electronic file may not be altered in any way. The author(s) of this article is/are permitted to use this PDF file to generate printed copies to be used by way of offprints, for their personal use only. Permission is granted by the publishers to post this file on a closed server which is accessible only to members (students and faculty) of the author's/s' institute. It is not permitted to post this PDF on the internet, or to share it on sites such as Mendeley, ResearchGate, Academia.edu. Please see our rights policy on https://benjamins.com/content/customers/rights For any other use of this material prior written permission should be obtained from the publishers or through the Copyright Clearance Center (for USA: www.copyright.com). Please contact [email protected] or consult our website: www.benjamins.com Data transparency and citation in the journal Gesture Lauren Gawne,1 Chelsea Krajcik,2 Helene N. Andreassen,3 Andrea L. Berez-Kroeker,4 and Barbara F. Kelly5 1 La Trobe University | 2 SOAS University of London | 3 UiT The Arctic University of Norway | 4 University of Hawai'i at Mānoa | 5 University of Melbourne Data is central to scholarly research, but the nature and location of data used is often under-reported in research publications. Greater transparency and citation of data have positive effects for the culture of research. This article presents the results of a survey of data citation in six years of articles published in the journal Gesture (12.1–17.2). Gesture researchers draw on a broad range of data types, but the source and location of data are often not disclosed in publications. There is also still a strong research focus on only a small range of the world’s languages and their linguistic diversity. Pub- lished papers rarely cite back to the primary data, unless it is already pub- lished. We discuss both the implications of these findings and the ways that scholars in the field of gesture studies can build a positive culture around open data. Keywords: gesture studies, data citation, open data, data management Introduction Gesture studies is a field founded on an empirical research method; our under- standing of gesture is based on evidence from data which is analysed and dissem- inated in research publications. Data is central to the formulation of analysis, but it is rarely presented in a way that is transparent to the reader.The transparency of data can refer to a number of features. These festures include how well the data is described in a research article, and whether the data is accessible in its entirety, or as a subset of specific examples, or has access restrictions. Transparency also includes citing the data to varying levels of granularity, directing the reader to a whole cor- pus, or to specific examples within that collection.There are many advantages to Additional material available from https://doi.org/10.1075/gest.00034.gaw.additional https://doi.org/10.1075/gest.00034.gaw Gesture 18:1 (2019), pp. 83–109. issn 1568-1475 | e‑issn 1569-9773 © John Benjamins Publishing Company 84 Lauren Gawne et al. having greater transparency of data in research practice – for authors, readers, and the field as a whole.These include heightened professional valuation of data collec- tion and sharing (Haspelmath & Michaelis, 2014; Thieberger,Margetts,Morey,& Musgrave, 2016) and greater accountability in research by facilitating access to the underlying data and methods (Gezelter, 2009). In order to best understand where the field of gesture studies is heading with regards to the use of data, we seek to understand the current state of practice.To do this, we conducted a six year survey of research publications in the journal Gesture, from 12.1 (in 2012) to 17.2 (in 2018). This survey examines how researchers describe the source and location of their data, and whether they cite examples back to the primary source.We also look at the types of data and the languages that researchers in gesture studies are working with, to better understand the support that will be needed to continue to develop a culture of research data transparency. While researchers in this field draw on a broad range of data types, the nature of this data is rarely made clear in publications. This has implications for the future progress of research.We discuss the results of our survey in light of the broader ‘open access’ movement, as well as the specific ethical implications of working with gestural, and particularly video, data.We also discuss the results in light of the move by Gesture to require greater transparency in data reporting. Background In the field of gesture studies, perhaps more than any other field in human com- munication, the means by which data is collected and analysed becomes crucial to the development and interrogation of theories underpinning the frameworks of data analysis. Gesture research draws on a range of different methodologies for analysing multimodality, particularly manual gestures and gaze.In early studies gestures were characterised as relatively static visual signs rather than dynamic signs changing across space and time (e.g., de Jorio, 1832; Morris,Collett,Marsh, &O’Shaughnessay, 1979). Thanks to affordable video capture and computers for analysis, recent research tends towards more empirical studies presenting transcribed, coded, and analysed gestures and affiliated spoken language.These empirical methods and analytic approaches yield ideal data sets for the replicabil- ity and reproducibility of findings. Gesture studies has a strong history of qualita- tive and quantitative research that spans multiple research fields. One thing that links all research in this area is a clear acknowledgement of the role of primary data in shaping our understanding of the form of gesture and its role in communi- cation.The discipline-spanning nature of gesture studies means that as a field we need to consider the multiple ways in which data transparency can lead to subse- quent research. © 2019. John Benjamins Publishing Company All rights reserved Data transparency and citation in the journal Gesture 85 Replicability and reproducibility have each received a good deal of attention in the social sciences lately, especially from those interested in the Open Access and Open Science initiatives (Buckheit & Donoho, 1995; de Leeuw, 2001; Donoho, 2010; Gawne & Styles, forthcoming, inter alia). While these terms may seem inter- changeable, the differences between them are crucial to the future of the language sciences. Replicability is probably the more widely familiar of the two concepts and is one that has underpinned the scientific process for a long time.Replicable studies are those studies that are created, executed, and subsequently described in such a way that another researcher could recreate the study down to the small- est detail.The results of this replicated study would either confirm eth previous results – lending them credence – or disconfirm emth .The aim of replicability is to ensure some level of scientific rigor in the research process, as well as to provide a mechanism by which results can be “checked” by those with a healthy degree of skepticism.Granted, it may not be enjoyable having ones research disconfirmed, but that is part of doing good science, and it says something positive about our methods that they were replicable in the first place. Replicability is the standard for scientific studies in which variables can be carefully controlled, such as in laboratory experiments. However, a great deal of science deals with data that is a little more “wild”(cf. the 2011 special issue of Sci- ence on reproducibility edited by Jasny,Chin, Chong,&Vignieri). This includes the behavioral data that is the basis of language-based research in many disci- plines (Berez-Kroeker, Gawne, et al., 2018). It is nearly impossible to create lan- guage studies that are truly replicable in the original sense of the word because it is very difficult to control for every factor that leads to the use of a particular word or gesture in a given linguistic context (be it naturalistic, elicited, or exper- imental). Even in the most tightly controlled language experiments, it would be impossible to control for a subject’s previous experience with a particular sound, word, phrase, or gesture.In such situations, the notion of reproducibility becomes valuable: reproducible research, therefore, is research that facilitates access to not just the methods used in the study, but also to the data collected in the study, and the tools (software, scripts, etc.) used to collect and analyse it.Another researcher could then examine or even reanalyse the data to reach similar or different con- clusions. Thus, when replicability is impossible, reproducibility steps in to ensure a level of rigor and accountability in the scientific process. Research that is reproducible or replicable requires a high degree of trans- parency on the part of scientists who must effectively communicate to their audi- ences about every aspect of their methodology, from collection to processing to analysis. Doing so would allow someone else to recreate the original study to test if the original hypothesis and analysis is supported.Replicability further requires clear description of the location of the underlying data set and how one would gain access. © 2019. John Benjamins Publishing Company All rights reserved 86 Lauren Gawne et al. The Open Data